Settings Reference¶

Complete reference for all SerOps settings modules.

Location: Click your profile avatar in the top-right header

Personal preferences accessible from any page:

Setting	Options
Language	English, Français
Appearance	Light, Dark, System
Timezone	Select from dropdown (shows current time)

These settings are user-specific and stored in your browser.

General Settings (Organisation)¶

Location: Settings → General

Setting	Description
Organisation Name	Display name for your org
Organisation Logo	PNG, JPEG, GIF or WebP up to 2 MB, shown next to the organisation name in the sidebar. Stored in the database, resized to at most 256 px.
Website	Organisation website URL

AI Settings¶

Location: Settings → AI Providers

Model Selection¶

Setting	Options
Primary Model	Gemini, Claude, OpenAI, DeepSeek, Ollama (any registered `AIModelProvider`)
Fallback Model	Same list, plus `None` to disable fallback
Use Custom Keys	Yes/No

Provider Configuration¶

Manage AI providers and private API keys in Settings > AI Providers > Manage Providers. Each organisation's admin sets its own API key and, when needed, its own base URL (for example a local Ollama endpoint).

The provider list itself (adding a provider, Edit Config, deleting one) is shared by every organisation on the install. It can be changed by platform staff (a Django staff user), or by the admin when the install has a single organisation.

Adding/Editing a Provider¶

Click Add Provider or Edit Config.
Display Name: Name shown in dropdowns.
Adapter Type:
- Google Gemini: For Gemini models via the google-genai SDK.
- Anthropic Claude: Native adapter using the anthropic SDK. Supports prompt caching via cache_control:ephemeral on the batch-analysis system prefix.
- OpenAI Compatible: For OpenAI, DeepSeek, Ollama, Mistral, or any v1/chat/completions-compatible endpoint. Prompt caching on OpenAI + DeepSeek activates automatically when the system-role prefix crosses 1,024 tokens.
API Base URL: Optional for cloud providers, required for local/self-hosted (e.g., http://localhost:11434/v1).
Default Model: The model ID to send in requests (e.g., llama3, claude-3-sonnet).
Active: Toggle to enable/disable.

Connection Testing¶

Enter your API Key for a provider.
Click the "Eye" icon to verify input.
Click Save.
Click Test Connection to verify the key works.
- Success: Shows "Connected" badge.
- Failed: Check the error message (usually "Auth Error" or "Invalid Key").

Troubleshooting¶

"Auth Error": Usually means the stored API Key is invalid or expired. Delete the key and re-enter it.
"Encryption Reset": If the backend secret key changes, old keys become unreadable. You will need to re-enter them.
System Defaults: If no custom key is provided, the system may use environment variable defaults if configured.

Delivery Channels¶

Location: Alerts & routing → Delivery Channels

One row per notification destination, following the Grafana contact point model. Multiple channels of the same kind are supported.

Kind	Auth	Configuration
Slack	Incoming Webhook URL	Paste the webhook URL
Microsoft Teams	Incoming Webhook connector URL	Paste the connector URL
PagerDuty	Events API v2 integration key	Paste the integration key
OpsGenie	API key + region	Key plus `us` or `eu`
Email	None	Recipients in `extra_config.to`

Secrets are encrypted at rest and never returned by the API.

This page also holds Rotate ingest secret, the per-organisation bearer token that inbound Alertmanager, Grafana, and Datadog webhooks authenticate with. The plaintext is displayed once only.

See Notifications for the full guide.

Routing Rules¶

Location: Alerts & routing → Routing Rules

Decides which delivery channels fire for which incidents — the Alertmanager route table, flattened.

Field	Description
Minimum severity	`LOW`, `MEDIUM`, `HIGH`, `CRITICAL`, or blank for any
Environments	List of environment names, blank for all
Services	List of service names, blank for all
Event types	`batch_analyzed`, `alert_triggered`, `critical_issue`, blank for all
Targets	Which delivery channels this rule notifies
Priority	Evaluation order (informational — results are unioned)

All matcher axes are ANDed. Every matching rule contributes its targets; the union is de-duplicated so a channel is contacted once.

Warning

With no active routing rules, nothing is sent regardless of how many delivery channels exist.

Notification Preferences (Per User)¶

Location: Settings → Preferences

Personal delivery preferences — channels, severity filters, quiet hours, and digest frequency. These are per-user and independent of the organisation-wide Delivery Channels above.

See Notification Preferences.

Security Settings¶

Location: Settings → Security

Password Validation¶

SerOps uses a NIST 800-63B compliant password policy:

Setting	Value
Minimum Length	8 characters (fixed)
Maximum Length	128 characters
Complexity Requirements	None
Password Expiry	Never

Note: Password policy is fixed and cannot be configured via UI. Validation occurs during password creation/reset with real-time feedback.

Session Settings¶

Setting	Description
Session Timeout	30m / 1h / 4h / 8h / 24h / Never
Token TTL	30 days (auto-refresh on activity, at most once per hour)

SSO Configuration¶

Supported providers: - Keycloak - Okta
- Azure AD - Google Workspace

Team Settings¶

Location: Settings → Team

Roles¶

Viewer (read only), User (read, run analyses, edit projects, update issues) and Admin (everything in the organisation, including credentials, team and settings). The full list of permissions per role is on Team Management. The Settings menu only lists the pages your role can open.

Actions¶

Invite user
Change role
Remove user

Integration Settings¶

Location: Integrations (sidebar)

Manage global provider integrations:

Provider	Status
GitLab	Connected/Not configured
GitHub	Connected/Not configured
Azure DevOps	Connected/Not configured

Observability Configuration¶

Location: Integrations (sidebar)

Log, metric, and trace providers are configured alongside SCM and CI/CD providers on the Integrations page — there is no separate Observability page.

SerOps supports a hybrid model for Logs and Metrics storage (Internal vs External). See Observability Providers for details.

Default Provider Configuration¶

These environment variables configure the System Default providers for organisations that haven't set up their own integrations.

Variable	Description	Default
`ELASTICSEARCH_HOST`	Default Log Provider (ES)	`http://elasticsearch:9200`
`ELASTICSEARCH_USERNAME`	ES User	`elastic`
`ELASTICSEARCH_PASSWORD`	ES Password	`changeme`

Sentinel Settings¶

Location: Settings → Sentinel

Controls the Sentinel tiered intelligence system that evaluates batches before AI analysis.

Gating Mode¶

Setting	Description
Enable Active Mode	When ON, quiet batches skip AI analysis (saves cost). When OFF, shadow mode — all batches run AI.

Validate in Shadow Mode First

Start with shadow mode (default) to observe Sentinel decisions for 1-2 weeks before enabling active mode. Check the dashboard Sentinel Intelligence card to verify decisions are accurate.

Threshold Overrides¶

Fine-tune detection sensitivity. Set to 0 to use system defaults.

Setting	Default	Description
Error Spike Multiplier	2.0	How many times above baseline error rate triggers an alert. Lower = more sensitive.
Z-Score Threshold	3.0	Standard deviations from mean for memory/CPU anomaly detection. Lower = more sensitive.
Max Quiet Batches	12	Force AI analysis after this many consecutive quiet batches (safety net).

Tuning Tips

For critical production services, use a lower multiplier (e.g., 1.5) to catch smaller error spikes.
Keep max quiet batches at 12 or below to ensure periodic AI review.

See Sentinel How-To for the full guide.

Mail Server (SMTP)¶

SMTP is configured via environment variables (EMAIL_HOST, EMAIL_PORT, EMAIL_HOST_USER, EMAIL_HOST_PASSWORD, EMAIL_USE_TLS, DEFAULT_FROM_EMAIL). There is no UI editor — SerOps follows the Grafana pattern: change .env, restart the backend.

Verify a config end-to-end:

docker compose exec backend python manage.py check_smtp your-address@example.com

The command prints which values are currently in effect and sends a test message.

See the Configuration Reference for the full env-var list.

Removed in 2026-09-01

An earlier release exposed a Settings → Mail Server UI backed by a SystemMailSettings DB row. It was removed on 2026-09-01 to eliminate a per-instance lifecycle-email hijack surface. If you're upgrading and had SMTP configured through the UI, copy those values into your .env before applying migration 0013_remove_systemmailsettings.

Data Settings¶

Location: Settings → Data Management

Retention Policy¶

Setting	Options
Auto-Cleanup	Enabled/Disabled
Retention Days	7-365 days
Frequency	Daily/Weekly/Monthly

Batch Settings¶

Setting	Default
Generation Interval	10 minutes
Max Batch Size	1000 logs

Audit Log Settings¶

Location: Settings → Audit

Setting	Options
Retention Days	30/90/365/Forever
Export Format	CSV, JSON

Clusters¶

Location: Settings → Clusters

Register the Kubernetes clusters your projects run on. Clusters scope observability integrations and are one dimension of the deployment filter used across the Data and Issues pages.

See Clusters.

Traces¶

Location: Settings → Distributed Traces

Manage OTLP ingest keys for distributed tracing. Each key is shown once on creation and stored as a hash afterwards, following the same model as the alert ingest secret.

See Distributed Traces.